\begin{table}[h!]

\caption{\label{tab:lee_bounds}Lee bounds analysis results}
\centering
\begin{threeparttable}
\begin{tabular}[t]{lrrr}
\toprule
Outcome & Diff. & Lower bound & Upper bound\\
\midrule
Accuracy Discernment & 0.315 & 0.312 & 0.322\\
Awareness & -0.011 & -0.013 & -0.006\\
Engagement Attitude & 0.105 & 0.102 & 0.111\\
Engagement Behavior & 0.018 & 0.017 & 0.025\\
Health Preferences & 0.207 & 0.203 & 0.214\\
Sharing Discernment & 0.207 & 0.201 & 0.214\\
Source Discernment & 0.21 & 0.204 & 0.216\\
\bottomrule
\end{tabular}
\begin{tablenotes}[para]
\small
\item \textit{Note:} 
\item This table summarizes the difference in means for observed outcomes and Lee bounds estimated using a rank-based trimming procedure. The lower bound assumes that attrited treated units would have had the best outcomes and trims the top of the treated outcome distribution. The upper bound assumes they would have had the worst outcomes and trims the bottom. The trimming proportion equals the difference in response rates between the treatment and control groups.
\end{tablenotes}
\end{threeparttable}
\end{table}
